We prospectively studied 196 consecutive survivors of out-of-hospital ventricular fibrillation (VF) not associated with acute myocardial infarction and 46 consecutive, control patients without prior ventricular arrhythmias. Programmed stimulation included two extrastimuli (S3 protocol) in all patients and three extrastimuli (S4 protocol) in the last 140 study patients and in all control patients. Sustained ventricular tachycardia (VT) or VF was not induced in any control patient. In study patients, logistic regression identified two independent predictors of induced, sustained VT for both S3 and S4 protocols: prior spontaneous, sustained VT (37 patients; p less than or equal to .001) and prior myocardial infarction (113 patients; p = .005). With the S3 protocol, sustained VT was induced in 54% of patients with both prior myocardial infarction and prior sustained VT vs 4% without either; with the S4 protocol, sustained VT was induced in 91% vs 13%, respectively. Eighty-three percent of induced VT episodes had a cycle length less than 300 msec, and all required termination by cardioversion or pacing. VF was induced only in survivors of out-of-hospital VF without prior, spontaneous, sustained VT (S3 protocol, 9%; S4 protocol, 24%) but not in study patients with prior sustained VT (S3, p = .10; S4, p = .05) or control patients (S3, p = .06; S4, p = .01). The mean coupling intervals of extrastimuli that induced VF were not significantly different from the intervals that induced sustained VT.(ABSTRACT TRUNCATED AT 250 WORDS)